A Machine Learning Framework for Predicting Hospital Costs and Revenues Based on Healthcare Resources and Patient Waiting Times

Effective management of healthcare resources and patient waiting times is critical for operational efficiency and financial sustainability in healthcare institutions. This study proposes a machine learning (ML)-based framework for jointly estimating total cost and revenue by modeling the impact of five key healthcare resources (physicians, nurses, clerks, exam rooms, and triage areas) and patient waiting time. Three regression-based ML algorithms—Partial Least Squares (PLS), Random Forest (RF), and Gradient Boosting (GB)—were applied and compared using a simulated dataset. The GB algorithm demonstrated the best cost and revenue estimation performance, with R² values of 0.981 and the lowest MAPEs of 1.869% and 1.913%, respectively. The results indicate that non-linear models, such as GB, better capture the complex relationships between hospital operations and financial outcomes. This study highlights the potential of data-driven approaches to support strategic decision-making in hospital management, offering predictive accuracy and operational insight into resource allocation, cost containment, and revenue enhancement.